Memoryoracle

You'd reach for this when you need persistent memory across agent sessions. It bundles vector storage for semantic search, recall mechanisms to retrieve past context, summarization to compress conversation history, and redaction tools to scrub sensitive data. The ten tools cover the full lifecycle of agent memory management, from storing embeddings to pulling relevant context back when needed. Useful if you're building agents that need to remember previous interactions or maintain knowledge bases that grow over time. The streamable HTTP transport means you can integrate it remotely without running the server locally.